A Performance Evaluation of Multi-Programming Model on a Multicore System with Virtual Machines

被引:0
|
作者
Ueno, Hitoshi [1 ]
机构
[1] Cc Ures Off, Atsugi, Kanagawa, Japan
关键词
virtual machine; multicore; multi-programming; process; thread; Linux; performance evaluation; supervisor mode;
D O I
10.1109/MCSoC.2014.52
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
For a lot of equipment like industrial equipment, automotive or consumer electrical appliances, embedded computers are widely used. Many computers are embedded in the equipment because it has excellent features and their controls are more complex. Recently, general purpose microcomputer has been made to have multicore CPUs to achieve high performance. Multicore CPU is also expected to be used widely because of embedded computers, but it is not clear how to use the computer. Virtual machine technology is proposed and used as the solution to the problem, as it makes possible the integration of some control computers into a high-performance embedded computer. However, it is not simple for engineers to design the performance of the system because the characteristics of virtual machines are different compared with real machines. In this article, the performance characteristics of multicore embedded computers with virtual machines are analyzed, with an assumption that the control programs are developed by multi-programming model using process or thread of Linux OS. For the evaluations, a benchmark program is designed. The results show a performance difference of 5%-10% between process cases and thread cases, regardless whether real machine or virtual machine is used. In performance comparison of real machine and virtual machine, 13%-88% performance differences are realized, which depend on the number of CPU cores. Furthermore, some unprecedented phenomena in performance are found in cases where the number of CPU cores exceeds the number of parallel executions of process/thread. The measured results are useful and can serve as a guide in designing system performance for multicore embedded computers.
引用
收藏
页码:321 / 328
页数:8
相关论文
共 50 条
  • [1] OPTIMIZING PERFORMANCE IN A MULTI-PROGRAMMING SYSTEM
    HOWARD, PC
    [J]. DATAMATION, 1969, 15 (01): : 65 - &
  • [2] A MULTI-PROGRAMMING SYSTEM FOR A MICROCOMPUTER
    SMITH, A
    BELL, DH
    [J]. SOFTWARE-PRACTICE & EXPERIENCE, 1982, 12 (05): : 405 - 410
  • [3] A RAPID TURNAROUND MULTI-PROGRAMMING SYSTEM
    IRONS, ET
    [J]. COMMUNICATIONS OF THE ACM, 1965, 8 (03) : 152 - &
  • [4] RANDOM INJECTION CONTROL OF MULTI-PROGRAMMING IN VIRTUAL MEMORY
    GELENBE, E
    KURINCKX, A
    [J]. IEEE TRANSACTIONS ON SOFTWARE ENGINEERING, 1978, 4 (01) : 2 - 17
  • [5] IMPROVING CPU UTILIZATION IN A MULTI-PROGRAMMING SYSTEM
    LAZOS, C
    YANDLE, JR
    [J]. COMPUTER JOURNAL, 1979, 22 (03): : 203 - 205
  • [6] MODEL OF MULTI-PROGRAMMING COMPUTER RUNNING 2 TYPES OF PROGRAMS
    KOGAN, YA
    LITVIN, VG
    [J]. AVTOMATIKA I VYCHISLITELNAYA TEKHNIKA, 1978, (03): : 68 - 73
  • [7] Performance Evaluation of OpenMP Applications on Virtualized Multicore Machines
    Tao, Jie
    Fuerlinger, Karl
    Marten, Holger
    [J]. OPENMP IN THE PETASCALE ERA, (IWOMP 2011), 2011, 6665 : 138 - 150
  • [8] OMINBUS - A MULTI-PROGRAMMING EXECUTIVE SYSTEM FOR ZERO GRADIENT SYNCHROTRON CONTROL COMPUTER
    KNOTT, M
    BRESCIA, A
    VALENTE, A
    [J]. BULLETIN OF THE AMERICAN PHYSICAL SOCIETY, 1969, 14 (02): : 234 - &
  • [9] Evaluation of virtual machine performance on NUMA multicore systems
    Cheng, Yuxia
    Chen, Wenzhi
    [J]. 2013 EIGHTH INTERNATIONAL CONFERENCE ON P2P, PARALLEL, GRID, CLOUD AND INTERNET COMPUTING (3PGCIC 2013), 2013, : 136 - 143
  • [10] On the Performance and Energy Efficiency of the PGAS Programming Model on Multicore Architectures
    Lagraviere, Jeremie
    Langguth, Johannes
    Sourouri, Mohammed
    Ha, Phuong H.
    Cai, Xing
    [J]. 2016 INTERNATIONAL CONFERENCE ON HIGH PERFORMANCE COMPUTING & SIMULATION (HPCS 2016), 2016, : 800 - 807